Бидний тухай
Багш ажилтан
In this work, we consider a case study of inventory management in the petroleum import company 'Petrovis' in Mongolia. The company operates 15 warehouses and over 360 gas stations across the country and accounts for providing 30% of national petroleum usage. Based on the statistical data of daily demand, supply, and remaining stock amount at each warehouse for 3 years, we established a simulation model for the company's operation. Numerical experiments show that with the new inventory management the company is expected to save 500 thousand USD every year from some unnecessary 26% reduced level of warehouse inventory.
The eld of articial intelligence has been developing rapidly in the past thirty years and it has changed many industries around the world. In this study, we did a survey analysis on the applications of machine learning techniques in healthcare. Starting with categorizing scientic articles on which machine learning methods are used in medical sciences, we continued to analyze various application tools in practice that support medical decision- making with machine learning techniques.
In this work, we consider a case study of inventory management [1] in the petroleum import company ’Petrovis’ Co.LTD in Mongolia. Motivations for this study stem from applications where, due to current high budget for daily remained petroleum, the company is willing to use the saved budget for other business opportunities if possible. The objective is to minimize the total budget for petroleum reserves subject to no run out of stock which implies customer demand for petroleum is fulfilled without any operational problems. The company operates 19 warehouses and over 300 gas stations across the country and accounts for providing 30% of national petroleum usage. Based on the statistical data of daily demand, supply and remaining stock amount at each warehouse for 3 years, we established a simulation model for the company’s operation. Numerical experiments show that with the new inventory management the company is expected to save 500 thousand USD every year from some unnecessary 26% reduced level of warehouse inventory.
In this research, we have worked on creating heuristic algorithms designated to solve the school timetabling problem automatically or semi-automatically where some parts of the schedule can be inserted manually beforehand. This problem is traditionally formulated as a high dimensional integer programming problem which is not always computationally tractable, yet every semester hundreds of school manager face and approximately solve it to create the school timetabling. We designed a specific genetic algorithm with the fitness function that can bring all the necessary schedule requirements into one setting through the matrix calculation and numerically tested the algorithm for several real-life cases. The algorithm has now been incorporated into the Education Sector Information System, a country level platform for connecting, supporting and monitoring the national education sub-sectors, and is available to use.
In this work, we consider a case study of inventory management [1] in the petroleum import company ’Petrovis’ Co.LTD in Mongolia. Motivations for this study stem from applications where, due to current high budget for daily remained petroleum, the company is willing to use the saved budget for other business opportunities if possible. The objective is to minimize the total budget for petroleum reserves subject to no run out of stock which implies customer demand for petroleum is fulfilled without any operational problems. The company operates 19 warehouses and over 300 gas stations across the country and accounts for providing 30% of national petroleum usage. Based on the statistical data of daily demand, supply and remaining stock amount at each warehouse for 3 years, we established a simulation model for the company’s operation. Numerical experiments show that with the new inventory management the company is expected to save 500 thousand USD every year from some unnecessary 26% reduced level of warehouse inventory.
In this research, we have worked on creating heuristic algorithms designated to solve the school timetabling problem automatically or semi-automatically where some parts of the schedule can be inserted manually beforehand. This problem is traditionally formulated as a high dimensional integer programming problem which is not always computationally tractable, yet every semester hundreds of school manager face and approximately solve it to create the school timetabling. We designed a specific genetic algorithm with the fitness function that can bring all the necessary schedule requirements into one setting through the matrix calculation and numerically tested the algorithm for several real-life cases. The algorithm has now been incorporated into the Education Sector Information System, a country level platform for connecting, supporting and monitoring the national education sub-sectors, and is available to use.